On the syntax of light-headed noun phrases in the Old Irish glosses: dealing with missing data in corpus linguistics

The Old Irish glosses in contemporary manuscripts are the most reliable evidence for Old Irish syntax. These glosses convey discontinuous utterances that depend on the Latin text to which they are attached. One of the most obvious consequences of this discontinuous and textually dependent character is that the glosses very often convey what we could consider incomplete utterances, i.e. linguistic units that do not have the status of an independent clause or sentence.When it comes to the study of NPs and their syntax, we face cases in which a gloss consists only of a bare NP. This paper deals with these isolated NPs, in particular, with those that are introduced by the light heads intí, aní and a N, and defends the idea that these NPs constitute utterances that may be considered as complete, as against the initial impression, and therefore, that they may be considered for an inquiry on the syntax of Old Irish NPs.This paper covers the following points: a basic quantitative description of the distribution of isolated light-headed NPs and the introduction of the notion of ‘missing data’ (cf. Osborne 2013: 128‒129); a discussion of the notion of ‘free NPs’, i.e. apparently incomplete but meaningful linguistic units; an overview of the linguistic features of isolated NPs, in particular animacy, which is a statistically significant factor explaining the difference between isolated/dislocated light-headed NPs, which are more often animate, and intraclausal light-headed NPs, which are more often neuter; an outline of a wider investigation of all OIr. NPs in order to check if the above-mentioned statistical skew is in fact more general, i.e. that animate NPs are more likely in general to appear as free NPs.